In its final specification from October 1949 the Mark I stored data in one 40-bit number (the accumulator) or two 20-bit instruction registers, and had two 20-bit address modifier registers, called "B-lines", which could function either as index registers or as so-called base address registers. This is the earliest known implementation of such index/base registers – an important innovation in computer architecture, unknown in other machines until the emergence of second-generation computers (approximately 1955–1964).
The Mark I could perform 40-bit serial arithmetic, with hardware add, subtract and multiply and logical instructions. It used a single-address format order code with thirty function codes. Standard instruction time was 1.8 milliseconds, but multiplication was much slower.
For memory the Mark I used two Williams tubes, each storing 64 rows of 40 points, for a total of 128 words. 64 words was considered to be a single "page", so the system stored 4 pages. In addition to the tubes were two magnetic drums, which could store 64 pages (32 tracks) each. Instructions were loaded into the machine by using paper tape.
The Mark I was developed from the Small-Scale Experimental Machine (SSEM), nicknamed Baby, developed by Frederic C. Williams and Tom Kilburn, which ran its first successful program on June 21 1948. The Manchester Mark I led to the Ferranti Mark I, the world's first commercially available general-purpose computer.